Use sequential time series bucket IDs when possible

XMLWordPrintableJSON

    • Fully Compatible
    • v7.0, v6.0, v5.0
    • Execution NAMR Team 2023-07-24
    • 105
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None

      We should adopt a hybrid approach for generating Bucket OIDs by using a full 8-byte counter and only randomizing those bytes when we hit OID collisions. This should still reap the benefits of SERVER-78080 and address PERF-4352.

      Context: as a result of SERVER-78080 (fully randomizing the last 8 bytes of bucket OID values), we saw a significant regression detailed in PERF-4352.

            Assignee:
            Dan Larkin-York
            Reporter:
            Fausto Leyva (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated:
              Resolved: